This repository includes the code and data necessary to replicate the Machine Learning results presented in the research paper.
Meat consumption varies globally due to a complex interplay of factors including gender, generation, culture, religion, income, and education. Historically, meat has symbolized masculinity, prompting some men to consume larger quantities to affirm male-associated traits. In contrast, abundant choices in modern food systems have led younger generations to reduce meat intake, informed by nutritional and environmental awareness. Despite being among the world's leading meat consumer, Brazil exhibits consumption disparities stemming from food access inequalities. Yet, the specific influences of gender and age on Brazilian meat consumption patterns remain underexplored. This study leverages data from the 2017-2018 Household Budget Survey (Pesquisa de Orçamentos Familiares - POF) to examine how gender and generational cohorts correlate with meat consumption in Brazil. Through descriptive statistics and multiple linear regression analysis, we identify differences in meat consumption trends across genders and generations. Furthermore, we investigated the socioeconomic predictors of meat consumption using methods of Machine Learning (ML). Notably, there is a generational shift with lower beef consumption (13% for women and 17% for men) from the Silent Generation to Generation Z, alongside a consistent or higher intake of white and processed meats (72% increase for women and 92% for men). Our findings reveal that, on average, women consume up to 6 grams less meat daily than men, with young women showing the lowest meat consumption among all groups. This low consumption, however, does not appear to be directly linked to access difficulties. Advanced analyses suggest that the reduced intake of red meats may correlate with access challenges, while the choice for white meat over red meat in vulnerable socioeconomic groups also influences overall meat consumption metrics. This research underscores the necessity of considering diverse factors when designing campaigns aimed at reducing meat consumption. It is crucial that efforts towards lower meat consumption for environmental and health reasons do not exacerbate existing inequalities in meat access.
1 - Clone this project
2 - Install the environment
cd paper_gender_meat
conda env create --file environment.yaml
conda activate paper_gender_meat3 - Run the notebooks on the folder src/
TBD